PICO
The rise in climate and weather-related risks such as floods, droughts and landslides affect millions of people and their properties. Early Warning Systems (EWS) coupled with anticipatory actions, are instrumental in tackling these threats. Water, a central focus of Sustainable Development Goal (SDG) 6, is integral to climate action and influences many other SDGs, emphasizing the need for accurate water-related data. The United Nations launched the Early Warnings for All (EW4All) initiative in November 2022 to ensure global EWS coverage. The quantity and quality hydrological data is critical for effective EWS and climate resilience. Moreover, the existence of different hydrological data from different sources, especially from non-traditional sources like machine learning (ML) and artificial intelligence (AI), remain underutilized by National Hydrological Services (NHS) and other users.
Accessing and processing hydrological data is often challenging due to its heterogeneity, necessitating significant effort to harmonize and integrate disparate sources. These barriers hinder effective water management and issuing early warnings in time. The WMO State of Global Water Resources report 20231 highlights the urgency of addressing data access and availability issues. Easy access to relevant data relies on machine-to-machine communication, which remains challenging for many agencies.
To address this, the WMO Hydrological Observing System (WHOS) provides an interoperable framework for data sharing, access and visibility using relevant technologies. It provides functionalities such as data publishing, standardization, visualization and linking global data centres and research communities. By integrating data from diverse sources, including ML/AI, global datasets, satellite observations, and individual researchers, WHOS enhances data visibility, fosters co-operation, and demonstrates the value of hydrological data collection. WHOS interfaces the big data and non-traditional data sources with NHS data systems using standardization and brokering approaches and open-source tools.
WHOS employs tools and standards like OSCAR, WHOS DAB, WIS2Box, Hydroserver2.0, HydroShare, WDE, WMDS, WCMP2.0, OGC WaterML2.0, etc. OSCAR serves as WMO’s official metadata repository, enabling users to query and view observing stations. The Discovery and Access Broker (DAB) standardizes and harmonizes data, while WIS2Box simplifies data publication and download. HydroServer2.0 is an open-source data management tool accessible to all users including LDCs and SIDS. Standards such as WCMP2.0 and OGC WaterML2.0 support unified data discovery and access. Additionally, Topic Hierarchy for hydrology enables users to receive real-time data notifications by subscribing to a Message Queuing Protocol broker.
The WHOS portal serves as a one stop data portal connecting hydrological data from countries, regional and basin organizations, research communities and global centres (IGRAC, GRDC, etc). Advances in AI, ML, satellite technology, and citizen science are resulting in vast amounts of data and WHOS integrates these data to support researchers, modelers and practitioners in water resource management.
WHOS provides interoperable data to EW4All, Water Resources Management and HydroSOS systems by bridging gaps between research and operational applications. It supports transboundary cooperation, joint data monitoring and sharing, while demonstrating the return on investment in hydrological data collection. By harmonizing and sharing hydrological data, WHOS is instrumental in mitigating hydrological hazards and fostering global collaboration.
1 https://library.wmo.int/records/item/69033-state-of-global-water-resources-report-2023
How to cite: Gebremichael, Y., Otieno, W., Boldrini, E., Korhonen, J., Berod, D., and Mones, J.: WHOS technologies: Connecting the dots for Advancing Global Hydrological Data Exchange and Harmonization, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-21408, https://doi.org/10.5194/egusphere-egu25-21408, 2025.
Since 1989, the Global Precipitation Climatology Centre (GPCC) provides several globally gridded precipitation analyses based on in situ rain gauge measurements. The underlying precipitation database is mainly made up of contributions from the national meteorological and hydrological services of around 190 countries worldwide, but also from data collections of international projects. The database expands continuously, regarding the number of stations and length of timeseries. All incoming data (metadata and observations) undergo a semi-automatic quality control to ensure a high quality of GPCC’s data sets.
The GPCC provides data sets that cover very long time periods of 40 up to 130 years (‘Full Data Daily’, ‘Full Data Monthly’, ‘Climatology’, ‘HOMPRA-Europe’), as well as near-real time analyses ('First Guess Monthly', 'First Guess Daily', 'Monitoring Product', ‘Provisional Daily Precipitation Analysis’ and 'GPCC Drought Index'). In 2024, GPCC released the second version of HOMPRA Europe (Homogenized Precipitation Analysis for Europe Version 2), presenting a unique homogenized data set based on thousands of stations from 1951 to 2015 for Europe. In 2025, the GPCC will release updates of two particularly valuable data sets: Climatology (1951-2020) and Full Data Monthly (1891-2024) & Daily (1982-2024). These data sets are based on the complete database of the GPCC (up to 126’000 stations) and all data undergo an extensive quality control. In addition, some major changes to improve the quality of the data sets are planned, like assessing the interpolation scheme and merging the Full Data Monthly with the Monitoring Product.
All gridded data sets presented are freely available in netcdf format on the GPCC website https://gpcc.dwd.de and referenced by a digital object identifier (DOI). The site also provides an overview of all data sets, as well as a detailed description and further references for each data set.
How to cite: Schirmeister, Z. L., Ziese, M., Rustemeier, E., Finger, P., Heller, A., Schulze, R., Zepperitz, M., Fränkling, S., Jahn, M., and Breidenbach, J. N.: New products of the Global Precipitation Climatology Centre, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-9546, https://doi.org/10.5194/egusphere-egu25-9546, 2025.
Together with National Geographic and ESRI, we have developed the World Water Map: insights (https://livingatlas.arcgis.com/wwm-insights/) which provides modelled historical, current, and projected data of freshwater availability by source and use by sector to inform research, conservation, policy, and journalism.
This World Water Map insights provides robust open-source data sets and visualization of freshwater supply and demand that can be queried by source, sector, region and country, and by past, present, and projected future climate models. Users can gather data and create visualizations that inform research and tell stories of freshwater use, investigate what’s driving demand, and inspire sustainable adaptations and solutions.
The ‘insights Version’ compiles 40+ years of hydrological data from the global hydrological model PCR-GLOBWB 2, a grid-based global hydrology and water resources model and GRACE satellite data to chart global freshwater supply and demand to a scale of 100 square kilometres.
This service allows users to make data summaries, obtain time series data and acquire direct access to state-of-the-art climate projections and large-scale hydrological data that is otherwise not accessible for a larger scientific and policy community.
How to cite: Wanders, N., Leijnse, M., Droppers, B., Paez Trujillo, A., and Bierkens, M. F. P.: The World Water Map Insights: visualizing global water demand and supply, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7222, https://doi.org/10.5194/egusphere-egu25-7222, 2025.
In recent years, there has been a significant increase in the release of datasets across various domains, including water resources. This surge is driven by advancements in computational and storage technologies, as well as the growing need to develop robust, accurate data-driven solutions to address challenges such as climate change, water scarcity, and environmental pollution. As a result, a wealth of national and global spatio-temporal datasets has become freely accessible online. These datasets are invaluable for applications like flood forecasting, climate change analysis, aquatic ecosystem management, improving drinking water safety, and optimizing wastewater treatment processes.
Despite the availability of these datasets, importing them into Python remains cumbersome. Researchers must often sift through multiple sources, including search engines, GitHub repositories, and various websites, to locate the necessary data. The diversity of data providers means datasets are frequently presented in inconsistent units and stored in varying formats. Additionally, many datasets require extensive preprocessing before they can be used for analysis or modeling. This makes acquiring, cleaning, organizing, and managing data a complex task requiring advanced data handling skills.
These challenges highlight the need for a unified, consistent, automated, and reusable framework for extracting hydrological and environmental data. The water-datasets package addresses this gap by leveraging data-handling tools such as Pandas, NumPy, xarray, and Shapely to offer a streamlined workflow for automatic data extraction from multiple sources in various formats.
hydro-harmony is a Python package designed for the automated downloading, parsing, cleaning, and harmonization of freely available water resource datasets related to rainfall-runoff processes, surface water quality, and wastewater treatment. The package currently supports 66 datasets, downloading and transforming raw data into consistent, easy-to-use analysis-ready data. This allows users to directly access and utilize the data without labor-intensive and time-consuming preprocessing.
The package comprises three submodules, each representing a different type of water resource data: `rr` for rainfall-runoff processes, `wq` for surface water quality, and `wwt` for wastewater treatment. The rr submodule offers data for 47,716 catchments worldwide, encompassing both dynamic and static features for each catchment. The dynamic features consist of observed streamflow and meteorological time series, averaged over the catchment area, available at daily or hourly time steps. Static features include constant parameters such as land use, soil, topography, and other physiographical characteristics, along with catchment boundaries. This submodule not only provides access to established rainfall-runoff datasets such as CAMELS and LamaH but also introduces new datasets compiled for the first time from publicly accessible online data sources. The `wq` submodule offers access to 16 surface water quality datasets, each containing various water quality parameters measured across different spaces and times. The `wwt` submodule provides access to 22,201 experimental measurements related to wastewater treatment techniques such as adsorption, photocatalysis, and sonolysis.
The development of water-datasets was inspired by the growing availability of diverse water resource datasets in recent years. As a community-driven project, the codebase is structured to allow contributors to easily add new datasets, ensuring the package continues to expand and evolve to meet future needs.
How to cite: Iftikhar, S., Abbas, A., and Beck, H.: AquaFetch: A Unified Python Interface for Water Resource Dataset Acquisition and Harmonization, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19958, https://doi.org/10.5194/egusphere-egu25-19958, 2025.
In the era of large-sample hydrology (LSH), there is still a lack in the availability of consistent data related to water quality. To address this gap, we introduce CAMELS-CH-Chem, a dataset inspired by the recently published CAMELS-Chem for the contiguous United States. CAMELS-CH-Chem extends CAMELS-CH (Catchment Attributes and Meteorology for Large-sample Studies in Switzerland) by integrating stream water chemical parameters and atmospheric deposition data for 115 monitoring stations across Switzerland. Spanning the same period as the CAMELS-CH 1981–2020, with consistent identifiers, it enables seamless integration with the original hydro-meteorological and landscape attribute data. The dataset primarily encompasses data from the Swiss Federal Office for the Environment. It includes time series of over 20 stream water chemistry constituents, covering both field and laboratory data on water temperature, dissolved oxygen, pH, and electrical conductivity both at hourly and daily time resolution; together with measurements of alkalinity, ammonium, Ca, Cl, dissolved organic carbon (DOC), dissolved reactive phosphorus (DRP), total organic carbon (TOC), HCO3, K, Mg, Na, total hardness, total dissolved nitrogen, total organic nitrogen, total phosphorus, NO3, NO2, Si, and SO4. The dataset also includes monthly time-series of stream isotope data (deuterium and oxygen-18), annual atmospheric deposition concentrations for NO3, NH4, NH3, NO2 and total inorganic nitrogen, and aggregated livestock density information. Switzerland, often referred to as the 'water tower of Europe,' offers a uniquely diverse setting for hydrological research, characterized by its varied climatic, topographic, and anthropogenic influences. This diversity, combined with the rapid changes driven by climate change, makes Swiss catchments and landscapes an interesting natural laboratory for studying evolving water systems. CAMELS-CH-Chem offers the opportunity to combine an extensive catchment characteristics and streamflow dataset with a detailed set of water quality parameters, facilitating new advances for LSH research. By providing the chance to enhance process-based understanding of the water cycle, this dataset supports studies that integrate both quantity and quality aspects of hydrological systems. To our knowledge, CAMELS-CH-Chem is only the second CAMELS dataset to incorporate such an extension. We anticipate that its release will inspire the development of similar datasets worldwide.
How to cite: Schönenberger, U., M. do Nascimento, T. V., Pool, S., Siber, R., Kauzlaric, M., Horton, P., Höge, M., Floriancic, M. G., Staudinger, M., Storck, F., Rinta, P., Seibert, J., and Fenicia, F.: Swiss water quality: extending CAMELS-CH with data on isotopes, water quality and atmospheric chemistry, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1208, https://doi.org/10.5194/egusphere-egu25-1208, 2025.
Historically, hydrologic studies have focused on one or a small number of basins, in many cases these studies were limited by data availability and computational resources. The increased availability of large hydrological datasets in the last 20 years, such as gridded meteorological data sets and streamflow timeseries, and increased computing resources have empowered large-sample hydrology studies having accessible and high-quality large datasets available to the science community to facilitate the evaluation of hydrologic processes and prediction questions. To support modeling and climate research efforts in the Nordics the CAMELS (Catchment Attributes and Meteorology for Large-sample Studies)-Nordic was collected and processed from multiple sources and databases into a coherent dataset for the entirety of Norway and Sweden. CAMELS-Nordic combines not only meteorological and hydrological, but also topography, climate, streamflow, land cover, soil, and geology data with python tools to update time series automatically. The development of the data package takes advantage of high-quality and freely available data from various Norwegian, Swedish, and European agencies. It includes: (1) daily forcing data (e.g. observations, interpolations, and modeled data) for catchments located in Norway and Sweden; (2) daily streamflow data; (3) digital elevation model; (4) catchment properties (size, location, elevation, and catchment files); (5) landcover; and (6) soil type data. Dataset time series span 1980 to 2022.
How to cite: Valseth, K., Valnes, L., Lappegard, G., Silantyeva, O., and Mardal, K.-A.: Development of CAMELS-Nordic, a large-scale hydrometeorological and catchment properties dataset for Norway and Sweden, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10411, https://doi.org/10.5194/egusphere-egu25-10411, 2025.
Initially limited by data availability and computing resources, hydrological modelling has advanced significantly with efforts to make test catchments widely accessible, by sharing hydrological data, like it was made by the landmark MOPEX initiative (Schaake et al., 2006). In France, the CAMELS-FR dataset (Delaigue et al., 2024c, 2024d) was developed to support large-scale hydrological studies, gathering near-natural catchments, in line with the CAMELS dataset already built in various countries (see e.g. Addor et al., 2017).
The CAMELS-FR dataset features daily hydroclimatic time series, also aggregated at the monthly and yearly time steps. In addition, the dataset includes catchment-specific attributes covering location, topography, climatic indices, gauging characteristics, hydrological signatures, hydrogeology, geology, soil characteristics, land cover, and level of human influences. The criteria for catchment selection were based on data availability, low-level of regulation, consistency of catchment area estimates, and data quality.
The first version of the CAMELS-FR dataset comprises data from 654 catchments across France, with time series spanning from 1970 to 2021, and 255 attributes organized into 10 classes. These catchments encompass a wide range of hydroclimatic contexts, from snow and groundwater-dominated to Mediterranean climates.
The CAMELS-FR dataset is complemented by graphical fact sheets (Delaigue et al., 2024a) that provide static summaries of hydroclimatic, topographical, hydrogeological, and land cover data, as well as dynamic graphs of hydroclimatic time series (Delaigue et al., 2024b) for interactive analysis.
Designed as a "living" dataset, CAMELS-FR will undergo updates to extend time series, correct streamflow values, and add new catchments, including overseas territories. Future versions may include data at finer temporal and spatial resolutions. An extension into the global Caravan initiative (Kratzert et al., 2023) is also planned.
References
Addor, N., Newman, A. J., Mizukami, N., and Clark, M. P.: The CAMELS data set: catchment attributes and meteorology for large-sample studies, Hydrology and Earth System Sciences, 21, 5293–5313, https://doi.org/10.5194/hess-21-5293-2017, 2017.
Delaigue, O., Brigode, P., Lobligeois, F., Bourgin, P.-Y., and Guimarães, G. M.: CAMELS-FR graphical fact sheets, https://doi.org/10.57745/KK2SVJ, V1, 2024a.
Delaigue, O., Génot, B., and Guimarães, G. M.: CAMELS-FR time series dynamic graphs, https://doi.org/10.57745/HBQWP5, V1, 2024b.
Delaigue, O., Guimarães, G. M., Brigode, P., Génot, B., Perrin, C., and Andréassian, V.: CAMELS-FR dataset, https://doi.org/10.57745/WH7FJR, V1, 2024c.
Delaigue, O., Guimarães, G. M., Brigode, P., Génot, B., Perrin, C., Soubeyroux, J.-M., Janet, B., Addor, N., and Andréassian, V.: CAMELS-FR dataset: A large-sample hydroclimatic dataset for France to explore hydrological diversity and support model benchmarking, Earth Syst. Sci. Data Discuss. [preprint], https://doi.org/10.5194/essd-2024-415, in review, 2024d.
Kratzert, F., Nearing, G., Addor, N., Erickson, T., Gauch, M., Gilon, O., Gudmundsson, L., Hassidim, A., Klotz, D., Nevo, S., Shalev, G., and Matias, Y.: Caravan - A global community dataset for large-sample hydrology, Scientific Data, 10, 61, https://doi.org/10.1038/s41597-023-01975-w, 2023.
Schaake, J., Cong, S., and Duan, Q.: The US mopex data set, IAHS Publication Series, 307, 9–28, https://www.osti.gov/biblio/899413, 2006.
How to cite: Delaigue, O., Guimarães, G. M., Brigode, P., Génot, B., Perrin, C., Soubeyroux, J.-M., Janet, B., Addor, N., and Andréassian, V.: CAMELS-FR dataset: A large-sample hydroclimatic dataset for France to explore hydrological diversity and support model benchmarking, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6659, https://doi.org/10.5194/egusphere-egu25-6659, 2025.
Large-scale datasets of hydrometeorological time series and catchment attributes are essential for advancing the understanding of hydrological processes, advancing hydrological model development, and enabling performance benchmarking. CAMELS (Catchment Attributes and Meteorology for Large-sample Studies) datasets have already been published for various regions worldwide covering a wide range of hydrometeorology and physiography.
We introduce a CAMELS-style dataset for Denmark (CAMELS-DK) containing predominantly lowland, groundwater-influenced, and small-sized catchments. With respect to already published CAMELS datasets, we see this as a valuable extension that enlarges the variability of catchments.
Moreover, this is the first CAMELS dataset to include both gauged and ungauged catchments as well as detailed groundwater information. CAMELS-DK comprises dynamic and static variables for 3,330 catchments across Denmark, derived from diverse hydrogeological datasets, meteorological observations, and simulated variables provided by the National Hydrological Model of Denmark. From the latter, a comprehensive list of simulated groundwater related variables like phreatic depth or groundwater-surface water interactions, are included. Streamflow observations are available for 304 catchments, while simulated streamflow data are provided for a total of 3,330 catchments. The dataset spans 30 years (1989–2019) at a daily temporal resolution. Additionally, the dataset includes variables capturing human impacts on Denmark's water resources, such as groundwater abstraction and irrigation.
By providing streamflow at almost full spatial coverage of Denmark, and not being limited to gauged sites, along with various simulation outputs from a distributed, process-based hydrological model, CAMELS-DK significantly enhances the utility of CAMELS datasets. This includes supporting the development of data-driven and hybrid/physically informed modeling frameworks.
The dataset is accessible via Koch et al. (2024) and the paper describing the dataset is currently under review (Liu et al., 2024).
Koch, J., Liu, J., Stisen, S., Troldborg, L., Højberg, A. L., Thodsen, H., Hansen, M. F. T., and Schneider, R. J. M.: CAMELSDK: Hydrometeorological Time Series and Landscape Attributes for 3330 Catchments in Denmark, https://doi.org/doi:10.22008/FK2/AZXSYP.
Liu, J., Koch, J., Stisen, S., Troldborg, L., Højberg, A. L., Thodsen, H., Hansen, M. F. T., and Schneider, R. J. M.: CAMELS-DK: Hydrometeorological Time Series and Landscape Attributes for 3330 Catchments in Denmark, Earth Syst. Sci. Data Discuss. [preprint], https://doi.org/10.5194/essd-2024-292, in review, 2024.
How to cite: Koch, J., Liu, J., Stisen, S., Troldborg, L., Højberg, A., Thodsen, H., Hansen, M., and Schneider, R.: CAMELS-DK: Hydrometeorological Time Series and Landscape Attributes for 3330 Danish Catchments with Streamflow Observations from 304 Gauged Stations , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6750, https://doi.org/10.5194/egusphere-egu25-6750, 2025.
The first large-sample catchment hydrology dataset for Aotearoa New Zealand, the Catchment Attributes and Meteorology for Large-Sample Studies—New Zealand (CAMELS-NZ), provides hourly hydrometeorological time series and detailed landscape attributes for 373 catchments across New Zealand. Spanning over the years 1972 to 2024, this dataset includes hourly records of streamflow, precipitation, temperature, and potential evapotranspiration. CAMELS-NZ offers a detailed set of catchment attributes that quantify physical characteristics such as land cover, soil properties, geology, topography, and human impacts. CAMELS-NZ integrates high-resolution time series data with static catchment characteristics, enabling the study of fast-rising rivers common in New Zealand. This dataset supports a wide range of hydrological research applications, including model development and climate impact assessments, prediction in ungauged basins and large-sample comparative studies. We include anthropogenic attributes on the presence of abstractions, dams, quality of rating and influences such as groundwater, snow or ephemeral rivers. CAMELS-NZ offers an opportunity to study hydrological processes in volcanic and alpine environments, while filling a critical gap of data in the Pacific region.
How to cite: Fischer, S., Bushra, S., Shakya, J., Cattoen-Gilbert, C., and Pahlow, M.: CAMELS-NZ: Hydrometeorological times series and landscape attributes for catchments in New Zealand, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1619, https://doi.org/10.5194/egusphere-egu25-1619, 2025.
Large-sample hydrological datasets containing data for tens to thousands of catchments are invaluable for hydrological process understanding and modelling. CAMELS (Catchment Attributes and MEteorology for Large-sample Studies) datasets are a family of large-sample hydrology datasets that contain hydro-meteorological timeseries, catchment attributes and boundaries for large-samples of catchments for specific countries or regions. CAMELS-GB was the first large-sample, open access data for Great Britain, consisting of hydro-meteorological catchment time series, catchment attributes (describing topography, climate, hydrology, land cover, soils, hydrogeology, and human influences), and catchment boundaries for 671 catchments.
While CAMELS-GB, released in 2020, is a valuable dataset, there are important gaps in the current dataset. Firstly, CAMELS-GB only contains daily hydro-meteorological timeseries, when sub-daily timeseries is often needed for flood characterisation in small catchments across Great Britain. Secondly, CAMELS-GB only contains static catchment attributes (i.e. one snapshot of a geophysical property in time) which makes it challenging to use for trend analyses. Thirdly, groundwater is an important resource in Great Britain, yet no groundwater level timeseries are available in CAMELS-GB.
Here, we present the second version of CAMELS-GB which contains new datasets including hourly hydro-meteorological timeseries, groundwater level timeseries, dynamic catchment attributes characterising changes in land cover and static catchment attributes characterising groundwater timeseries and reservoirs. We update the existing data in CAMELS-GB to lengthen the timeseries of the daily hydro-meteorological timeseries and include the latest rainfall and PET data for Great Britain. The data will be made open access and available on the Environmental Information Data Centre.
How to cite: Coxon, G., Zheng, Y., Barbedo, R., Fileni, F., Fowler, H., Fry, M., Green, A., Harfoot, H., Lewis, E., Qiu, X., Salwey, S., and Wendt, D.: CAMELS-GB v2: hydrometeorological time series and landscape attributes for 671 catchments in Great Britain, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4371, https://doi.org/10.5194/egusphere-egu25-4371, 2025.
About two years ago, we started the Caravan community dataset. The idea behind was two-fold:
1) Caravan standardizes the streamflow data from different regional large-sample hydrology datasets (e.g. various CAMELS datasets) and combines them with data from globally available data sources.
2) All data in Caravan, besides streamflow, is derived from Google Earth Engine, with code that has been made publicly available (https://github.com/kratzert/Caravan/ ) , allowing anyone to extend the dataset to new regions.
Additionally, the dataset structure allows for easy integration of what we call “community extensions” and so far, six different community extensions (https://github.com/kratzert/Caravan/discussions/10 ) have been made available, extending Caravan to a total of 22494 gauges.
With this submission, we want to present a new kind of extension to the Caravan project, which does not add new basins (i.e. streamflow data) but rather adds additional weather data for all existing basins. More specifically, we add three additional precipitation nowcast products (CPC, IMERG Early v.0.7, and CHIRPS), and three weather forecast products (ECMWF IFS HRES, GraphCast, and CHIRPS-GEFS). For the ECMWF IFS forecast data, as well as for GraphCast, we include not only precipitation but several land surface variables.
Since not all of this data is available on Earth Engine, we process this data for all existing Caravan gauges, including all extensions. In agreement with the existing data in Caravan, we spatially average all weather data across the catchment area and aggregate to daily resolution. However, since not all data can be easily shifted to local time (as with the original ERA5-Land data in Caravan), we keep all weather products in UTC and therefore also include ERA5-Land in UTC for consistency.
To our knowledge, this extension to Caravan makes it the first large-sample hydrology dataset that includes real weather forecast data. We hope that this extension can be used to enable and empower hydrological research, specifically working on forecasting problems.
How to cite: Shalev, G. and Kratzert, F.: Extending Caravan with additional weather nowcast and forecast products, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-4768, https://doi.org/10.5194/egusphere-egu25-4768, 2025.
Italy, like many other Mediterranean countries, is increasingly facing meteo-hydrological hazards, with extreme weather events becoming more frequent and intense. These changing conditions pose serious risks to dam safety and call for a reassessment of spillway design floods to adapt to evolving hydrological scenarios.
Historical flood data may no longer reflect the full range of potential events, highlighting the urgent need for updated and accurate hydrological information. In response to this challenge, the Italian Large Dams Committee emphasized the importance of improving hydrological data at dam sites to improve flood management strategies.
In this context, we present a comprehensive dataset that include geometrical characteristics, as well as watershed-related features, for all 528 large dams in Italy. Freely accessible on Zenodo, this dataset represents the most extensive resource available on Italian dams, providing precious structural and environmental information to researchers, policy makers and stakeholders involved in water resource management and infrastructure planning.
The dataset presents detailed structural information about each dam, such as commissioning year, height, and type, alongside data on reservoir features like volume, surface area, and intended uses. Some of this information is sourced from the General Department of Dams and Hydro-Electrical Infrastructures. Notably, it also includes critical parameters such as reservoir surface area and the elevation of the maximum water level allowed, which are essential for assessing each dam’s capacity to mitigate flood peaks effectively.
On the other hand, key catchment characteristics, such as size, shape, slope, and land cover, are crucial for modeling flood scenarios and addressing these escalating risks. The database contains basin characteristics, including geomorphological, soil, land cover and climatic attributes, as well as basin boundaries, that are determined using standardized and uniform procedures, ensuring consistency throughout the country. Taking into account the availability of the "twin" dataset from Claps et al. (2024), a wide level of detail is therefore provided on about a thousand watersheds, all over Italy, including both dammed and gauged watersheds.
As climate change and water resource challenges intensify, a thorough understanding of our existing infrastructure becomes crucial. In this sense, this work can help to improve our capability to manage the complex interplay between dams and their hosting environment.
How to cite: Evangelista, G., Mazzoglio, P., Ganora, D., Pianigiani, F., and Claps, P.: A new resource on Italian large dams, their catchments, and key attributes, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15538, https://doi.org/10.5194/egusphere-egu25-15538, 2025.
Water scarcity in the Mediterranean region is increasing due to climate change and anthropogenic pressures. to This situation has intensified drought, reduced streamflow, and decreased soil moisture, putting additional pressure on groundwater to mitigate water stress. Despite the critical importance of groundwater use, there is a lack of centralized and detailed groundwater level data in the Mediterranean region, which is essential for sustainable water resources management. This study aims to establish a comprehensive, accurate and up-to-date groundwater level dataset in the Mediterranean region.
The dataset was primarily constructed using available nationwide observation wells from Mediterranean countries and was further enriched with additional research project-based observation wells. It includes groundwater level data with more than 15800 observation wells in Portugal, Spain, France, Italy, Greece, Türkiye, and Tunisia. The historical data (1900-2024) frequency varies from daily to weekly, monthly, bimonthly, and biannually, with the earliest records originating from France. The highest proportion of daily and weekly measurements also comes from wells in France, whereas 90% of the observations in Portugal and Spain were recorded at monthly or bimonthly intervals. The collected groundwater dataset will be presented and discussed. Consistent and detailed monitoring and sharing of groundwater level data are essential for sustainable water management, especially in areas known for water scarcity and rapidly changing climatic conditions.
Acknowledgment: This work was supported by the OurMED PRIMA Program project funded by the European Union’s Horizon 2020 research and innovation under grant agreement No. 2222.
How to cite: Jomaa, S., Chávez García Silva, R., Rouhani, A., Ben-Salem, N., K. Copty, N., Khlifi, S., Ben Nsir, S., A. Varouchakis, E., Rode, M., Elçi, A., Andrew Barry, D., and Gómez-Hernández, J. J.: Groundwater level observations dataset for the Mediterranean region, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17977, https://doi.org/10.5194/egusphere-egu25-17977, 2025.
Flood event analysis, using both rainfall and river flow data, enables understanding of how river catchments respond to different rainfall events under varying antecedent conditions, initial conditions, and rainfall event characteristics. (e.g. mean and peak rainfall intensity, initial flow rate, antecedent rainfall total). Large datasets of flood events, collected across many river catchments, can enhance our understanding of how catchment properties influence catchment response, and how changes in catchment and climate characteristics cause changes in rainfall-runoff relationships during flood events. These datasets can support the development of flood risk and flood response models, and increase resilience to extreme events.
The UKCEH Flood Event Data Suite consists of a procedure to identify paired rainfall-runoff events from paired time-series, the dataset of events identified when applied to open-access UK precipitation and flow data, and a database of rainfall statistics and runoff signatures derived from those events.
The procedure identifies paired rainfall-runoff events by starting at peak flow and extending forwards and backwards to common start and end times that encapsulate both complete rainfall and complete runoff events. Using open-access data, the procedure is intended to be re-run periodically as the input data sources are updated, with the resulting dataset versions made available through an interactive public portal hosted by the UK National River Flow Archive, and static “snapshots” released periodically through the UK Environmental Information Data Centre.
The dataset in its most recent form includes approximately 175,000 paired rainfall-runoff events extracted from 1200 gauged catchments over a long monitoring period (1990-2016 inclusive), for approximately 5.4 events per station per year, at a high time resolution consistent across all events (15 minutes for flow, 60 minutes for rainfall). Other key features include a high station density (approximately 1 per 200 km2), a wide range of catchment properties, including area (< 1 to approximately 10,000 km2), mean annual rainfall (approximately 500 to 3500 mm/year), baseflow index (approximately 0.2 to 1) and urbanization (0% to approximately 70%), long flow recessions captured, publicly available catchment shapefiles, allowing users to extract information from other spatial datasets using consistent catchment outlines, a single peer-reviewed source for all rainfall data, and (planned) wide availability through an open access repository and portal.
Analysis of a high-quality subset of the derived database of rainfall statistics and runoff signatures (677 catchments, ~7500 events) has identified that different catchment descriptors influence peak flow, total volume and rate-of-rise, changing with event rarity, correlations between peak flow and total volume increase with catchment size, and summer events typically have high rates-of-rise and volumes relative to baseflow, while spring events typically have the opposite.
How to cite: Vesuviano, G., Fry, M., Swain, O., Cooper, H., Fileni, F., Chevuturi, A., and Khamis, D.: The UKCEH Flood Event Data Suite, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7031, https://doi.org/10.5194/egusphere-egu25-7031, 2025.
The Flow Duration Curve (“FDC”) is one the most effective and practical tools in hydrological sciences that not only enhances the understanding of the hydrological process of basins but also assists in analyzing water availability and stream flow fluctuations. The most significant challenge hydrologists have to face is an absence or scarcity of flow data in ungauged basins, where direct measurements are not feasible. To tackle this issue, regionalization of FDCs has emerged as a persuasive method. This approach, while applying available information in different basins, allows for reconstructing the hydrologic response in unmonitored basins.
This study focuses on the regionalization of FDCs in Southern Italy, containing 114 hydrological stations located in the Regions of Campania, Basilicata, Calabria, and Puglia, whose surface accounts for about 20% of the Italian country. For these areas, the only available authoritative information consists of archival reports of daily discharge and monthly rainfall and runoff over the period 1924 to 1994, characterized by gaps, repetitions, and ambiguities. Also, the complete river network and the outer boundary of the primary basins are available. No precise information is available for the location of the monitored sections.
For the first step (completed), historical data on daily discharge and monthly rainfall and runoff from 1924 to 1994 were digitized from archival paper reports collected by the National Hydrological Service (now expired) for 114 monitored sections, and their reliability was verified. For all the catchments, a hydrologically connected Digital Elevation Model was created integrating the authoritative Italian 20m × 20m DTM with the abovementioned physiographic information: this allowed for accurate reconstruction of physical, hydrological, topographic, and morphologic features in the GIS environment. The database was further refined with land use/land cover data from the CORINE initiative, with geological information coming from local maps, and with ancillary variables such as the baseflow index.
Currently, efforts are in progress to identify the main dependencies between flow variables and covariates in the database, to find common patterns and recursive behaviors. However, the data mining process is deeply intertwined with the choice of the regionalization methodology, which drives the main parameters and quantities to be investigated. Some issues that are being explored at present comprise (but are not limited to): i) understanding differences, similarities, and potential of annual FDCs and total FDCs (i.e. obtained by year-by-year or full-period records respectively); ii) investigation about relevant percentiles of FDCs representing low, high and semi-perennial flows; iii) normalization of FDCs; iv) cluster analysis also relying on Artificial Intelligence.
At a more mature stage, research will allow us to compare the results of different regionalization techniques, including parametric and non-parametric approaches, statistical methodologies, and quantile-based techniques. A comprehensive assessment of the relationship between hydrological and physiographic features will promote the construction of predictive models for streamflow behavior in unmonitored basins, fostering more efficient water resource planning and management in data-constrained locations.
Keywords: Regionalization, Flow Duration Curve, Digitalization, Geographic Information System (GIS), Southern Italy
How to cite: Moradi, F., Padulano, R., and Delgiudice, G.: Regionalization of Flow Duration Curves in Southern Italy, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6922, https://doi.org/10.5194/egusphere-egu25-6922, 2025.
Complete hydrological time series are critical for effective water resource management, flood and drought forecasting, hydroelectric power optimization, irrigation planning, ecological preservation, and climate change impact assessments. However, significant data gaps in streamflow and water level observations, compounded by extreme hydroclimatic events and quality control issues, hinder accurate modeling and informed decision-making in Indian catchments. The current challenges are particularly pronounced in regions with high climatic variability, where missing data spans 6 to 12 months. To address this, we employed geomorphological, meteorological, and hydrological parameters in combination with the Random Forest method to gap-fill streamflow data at 352 stations across India, except the transboundary basins. To enhance model accuracy and training, we categorized stations into similar-behaving classes using a k-means clustering algorithm based on catchment characteristics. This clustering increased the availability of training data for machine learning models. Streamflow data from each class was trained with 80% of the available data and validated on the remaining 20%. Our results indicate that clustering significantly improves performance, with over 100 stations reporting a >25% increase in Nash-Sutcliffe Efficiency (NSE). Model performance was evaluated for continuous data gaps of 1 week, 1 month, 3 months, 6 months, and 1 year, revealing a decline in accuracy with longer gaps. Despite this, the mean NSE exceeded 0.85 across all clusters. The gap-filled datasets provide robust hydrographs, enabling precise streamflow variability modeling, climate-hydrology interaction evaluation, and improved water resource management strategies.
How to cite: Solanki, H. and Mishra, V.: Filling Streamflow Data Gaps in Indian Catchments Using Machine Learning, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15030, https://doi.org/10.5194/egusphere-egu25-15030, 2025.
Understanding how data uncertainties impact on our analyses is essential to draw the right conclusions about hydrological processes and their change in space and time. However, understanding the impact of data uncertainties is challenging when working with large numbers of catchments: data uncertainty estimates are rarely available from data providers and producing such estimates requires substantial efforts.
In the absence of such ‘hard’ information about data uncertainty, Westerberg and Karlsen (2024), suggested using ‘soft’ information to qualitatively estimate discharge data uncertainty and to summarise the soft information in a generalized perceptual model of uncertainty. They use three categories of soft information: station characteristics, climate and flow regime, and catchment characteristics. An example of soft information about the flow regime is that there are rare extreme high flows, this impedes high flow gauging and therefore increases high flow uncertainty. Another example of a climate characteristic is the presence of river ice-cover that increases low flow uncertainty.
In this study, we take the generalized perceptual model of discharge data uncertainty from Westerberg and Karlsen and translate it into catchment and station metadata from the Camels-GB dataset (Coxon et al., 2020) and the UK National River Flow Archive (https://nrfa.ceh.ac.uk). This enables us to evaluate how useful soft information is to identify stations with low and high data uncertainty by comparing it to the ‘hard’ uncertainty estimates at hourly and daily time scales from previous detailed stage–discharge rating curve uncertainty analyses (Coxon et al., 2015; Westerberg et al., 2016). We explore which metadata are most useful as soft information about high and low flow uncertainty respectively and recommend useful metadata on discharge data uncertainty to be included in large sample datasets.
Coxon, G., J. Freer, I. K. Westerberg, T. Wagener, R. Woods, and P. J. Smith (2015), A novel framework for discharge uncertainty quantification applied to 500 UK gauging stations, Water Resour. Res., 51, 5531–5546, doi:10.1002/2014WR016532.
Westerberg, I. K., T. Wagener, G. Coxon, H. K. McMillan, A. Castellarin, A. Montanari, and J. Freer (2016), Uncertainty in hydrological signatures for gauged and ungauged catchments, Water Resour. Res., 52, 1847–1865, doi:10.1002/2015WR017635.
Coxon, G., Addor, N., Bloomfield, J. P., Freer, J., Fry, M., Hannaford, J., Howden, N. J. K., Lane, R., Lewis, M., Robinson, E. L., Wagener, T., and Woods, R. (2020): CAMELS-GB: hydrometeorological time series and landscape attributes for 671 catchments in Great Britain, Earth Syst. Sci. Data, 12, 2459–2483, https://doi.org/10.5194/essd-12-2459-2020.
Westerberg, I. K., & Karlsen, R. H. (2024). Sharing perceptual models of uncertainty: On the use of soft information about discharge data. Hydrological Processes, 38(5), e15145. https://doi.org/10.1002/hyp.15145
How to cite: Westerberg, I. and Coxon, G.: Can we use soft information to estimate discharge data uncertainty for large samples of catchments?, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16412, https://doi.org/10.5194/egusphere-egu25-16412, 2025.
To say that high quality observatory datasets play an important role in hydrological research would be an understatement. Their accessibility is a crucial aspect in the scientific practice. When we talk about access of these datasets, we often hear one famous buzzword: FAIR - making data findable, accessible, interoperable and reusable.
But fulfilling the FAIR-principles requires the enrichment of our observatory datasets with comprehensive and consistent metadata. In particular, we need metadata about spatial and temporal context as well as environmental conditions, logger settings, sensor accuracy and resolution, as these parameters directly impact the usability of the data. Furthermore, the integration, management and export of all these information should be as user-friendly and consistent as possible so that researchers and instrument operators do not have to cope with complex metadata standards, terminologies and semantics.
In the DataHub initiative of the Helmholtz Centers of the Research Field Earth & Environment, we have hence developed the so-called Sensor Management System (SMS) as user-friendly one-stop-platform for collecting, managing and providing all senor-related metata in a homogenized and standardized way. Our system further supports the registration of devices and the generation of PIDs via B2INST, the documentation of changes on measurement setups, the linkage with data infrastructures as well as an API for machine-to-machine interaction, e.g., within data science applications.
The SMS has now reached a level of maturity that allows the full-fledged management of comprehensive sensor and measurement infrastructures, like the ones operated by the Hydrology-Section of the Helmholtz Centre for Geosciences. In this contribution, we therefore want to present the current state of our Sensor Management System, how it can simplify the management and ensure a sustainable and transparent operation of hydrological sensor systems and, finally, help hydrologists and instrument maintainers to make their research data FAIR.
How to cite: Lorenz, C., Brinckmann, N., Kuhnert, T., Günther, K., Thoss, H., Loup, U., and Schäfer, D.: The Helmholtz Earth & Environment Sensor Management System in hydrological sciences - A case study, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16357, https://doi.org/10.5194/egusphere-egu25-16357, 2025.
The Early Warning Data Store (EWDS), introduced by the European Centre for Medium-Range Weather Forecasts (ECMWF), represents a major advancement in the Copernicus Emergency Management Service (CEMS). Launched on 26 September 2024, the EWDS is a modern, user-focused system for hosting and disseminating CEMS-Flood datasets from the European and Global Flood Awareness Systems (EFAS and GloFAS), which provide vital data for flood forecasting, disaster preparedness, and water resource management.
In 2024, the Early Warning Data Store (EWDS) registered 18,003 users, with over than 20,138 completed requests for EFAS and GloFAS datasets from 42 distinct countries. The total data retrieved amounted to approximately 3,029.72 TB, distributed across different products. The EWDS hosts EFAS and GloFAS datasets in GRIB and NetCDF formats, including historical data, forecasts, reforecasts, seasonal forecasts, and seasonal reforecasts. Auxiliary datasets support flood forecasting and hydrological analysis. For EFAS and GloFAS, these include datasets related to upstream areas, elevation, soil characteristics, and flood thresholds.
Accessing these datasets is simplified through a modern web interface and an Open Geospatial Consortium (OGC)-compliant API, ensuring compatibility with diverse user needs. Following FAIR principles (Findable, Accessible, Interoperable, Reusable), the EWDS makes its data easy to find, access, and use across different platforms. Improvements to previous versions include flexible data download options and precise region-of-interest bounding box specifications. Supported by ECMWF's robust Meteorological Archival and Retrieval System (MARS) infrastructure, the EWDS ensures efficient data extraction and delivery, even for large-scale requests.
A key feature of the EWDS is its integration with Earthkit, ECMWF’s open-source Python project designed to simplify data workflows. Earthkit provides tools for data access, processing, analysis, and visualization, using libraries such as numpy, pandas, and matplotlib. Earthkit-Hydro, currently under development, will extend these capabilities, offering customized solutions for hydrological research and flood risk management. Additionally, there is comprehensive documentation and a user support and feedback service. This presentation will introduce the technological innovations of the EWDS, its user-focused capabilities, and its role in advancing global flood forecasting and risk management.
How to cite: Azhar, M., Prudhomme, C., Harrigan, S., Comyn-Platt, E., Morrison, O. M., Damasio da Costa, E., and Carton de Wiart, C.: Modernizing Flood Forecast Data Access with the CEMS Early Warning Data Store (EWDS), EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-21044, https://doi.org/10.5194/egusphere-egu25-21044, 2025.
Globally, access to hydrometric data with sufficient record length, quality, and geographical coverage to address research questions and manage freshwater systems remains a significant challenge. The UK National River Flow Archive (NRFA) oversees river flow data from over 1,600 locations across the UK. Data from almost 1,000 stations are acquired and displayed as ‘provisional’ in real-time, while the NRFA conducts a comprehensive update to the quality-controlled dataset annually. Upon submission, river flow records undergo both automated data screening and manual quality control by trained hydrologists to ensure the highest quality data are disseminated to the Archive’s broad user community, making it fit-for-purpose for various applications.
In the 1990s, increasing gaps in river flow records and declining data quality led to the introduction of a Service Level Agreement (SLA) in 2002, to safeguard the UK’s hydrometric network and its data. This paper presents the results from 20 years of applying the SLA system, analysing a set of quantifiable indicators of data quality, completeness and provision. The observed improvements underscore the advantages of the SLA in enhancing the reliability of the nationally archived river flow data. Furthermore, it serves as a model for quality assurance and performance measurement systems that can be adopted as best practice by other monitoring networks globally. These results also demonstrate a method of helping to ensure hydrological databases provide high-quality information to meet current and future research and water management needs.
How to cite: Suman, G., Turner, S., Muchan, K., Sefton, C., Kumar, A., Fry, M., Swain, O., Hannaford, J., and Tindall, I.: Analysis of the quality and completeness of UK river flow data - a long-term view, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3576, https://doi.org/10.5194/egusphere-egu25-3576, 2025.
In highly managed rivers where the presence of reservoirs has completely altered the natural regime of the river, returning to the historical natural flow regime is an increasingly distant option. However, managers and water authorities must establish criteria to ensure the ecological status of water bodies following the recommendations of the Water Framework Directive (WFD). Often integrating this objective competes with the rest of water uses, making maintaining a balance that brings on both human development and environmental conservation a challenge. This issue is especially crucial in unmonitored basins in which the natural hydrological regime conditions prior to alteration are unknown.
To face this challenge, interdisciplinary collaborations connecting stakeholders and research arise for both improving operational hydrology services and achieving science-informed policies. This work brings an example of that science-policy-practice nexus through a collaboration between university and business. The pilot area to host the experience is the Rivera de Huelva basin, located in the South of Spain, where a state-owned water company manages the three reservoirs which spatially and temporally allocate water resources in the basin. An approach combining historical streamflow and operation data and hydrological modelling will allow us to assess natural hydrological conditions in this unmonitored and regulated basin, as well as the definition of an environmental flow regime in drought and/or scarcity situations. Our outcomes will help reservoir managers to set the basis for the design of new minimum environmental flow rates which are founded on the principles of the natural flow regime paradigm.
Acknowledgments: This work has been funded by the project CONV 39-27 UCO-EMASESA, in the framework of the TED/934/2022-PCAU00006, funded by MITECO and by European Union NextGenerationEU/PTR.
How to cite: Contreras Arribas, E., Gómez Beas, R., Pimentel Leiva, R., Domínguez Romero, L., Escot Muñoz, C., and Polo Gómez, M. J.: Toward the return to the historical natural flow regime in highly managed rivers: a pilot experience in the South of Spain, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-20107, https://doi.org/10.5194/egusphere-egu25-20107, 2025.
Understanding hydrological systems is vital for addressing water-related challenges, including hydrological extremes such as floods and droughts. A complex interplay of climate and land-use changes shapes European river flow. While this interplay has been studied in specific subregions, a comprehensive analysis across continental to regional scales is yet to be unravelled.
In this study, we use the extensive European streamflow dataset, EStreams (do Nascimento et al., 2024), to investigate the relative influence of climate and landscape characteristics on streamflow behaviour across 7000 catchments.
To identify the primary drivers of streamflow behaviour, we first clustered catchments into 10 groups based solely on their hydrological signatures, deliberately excluding climate-related signatures to focus on hydrological similarity. This approach ensured that each cluster represented distinct patterns of streamflow behaviour.
Further, the drivers of these clusters were explored at both continental and regional scales. While climate emerged as the dominant driver of streamflow behaviour at the continental scale, a different pattern was observed within the clusters. By analysing regional-scale variability within each cluster, landscape characteristics—such as topography, geology, vegetation and soil properties—were found to play a larger role in shaping streamflow.
The relative contributions of climate and landscape characteristics were quantified using random forest models, applied separately to each cluster. These models revealed the relative importance of individual factors, offering insights into the nuanced controls of streamflow behaviour.
This analysis highlights three key findings. First, distinct clusters of hydrologically similar catchments can be identified across Europe using streamflow-based signatures alone. Second, climate characteristics are the primary drivers of streamflow behaviour at the continental scale. Third, dominant landscape characteristics are identifiable at the regional scale when accounting for within-cluster variability.
In conclusion, this study highlights the importance of multi-scale approaches to understanding hydrological systems. Using EStreams, a detailed perspective is offered on the interplay between climate and landscape in shaping European streamflow.
References
do Nascimento, T. V. M., Rudlang, J., Höge, M., van der Ent, R., Chappon, M., Seibert, J., Hrachowitz, M., & Fenicia, F. (2024). EStreams: An integrated dataset and catalogue of streamflow, hydro-climatic and landscape variables for Europe. Scientific Data, 11(1), 879. https://doi.org/10.1038/s41597-024-03706-1
How to cite: Rudlang, J. M., do Nascimento, T. V. M., van der Ent, R., Fenicia, F., and Hrachowitz, M.: Unravelling the Drivers: Spatial Insights into Climate and Landscape Influences on European Streamflow Behaviour, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11253, https://doi.org/10.5194/egusphere-egu25-11253, 2025.
The rainfall-runoff process at the catchment scale is governed by a complex interplay of physical and climatic driving mechanisms that vary in space and time. Although this variability makes it challenging to generalize catchment hydrological processes, classifying catchments in terms of their similarity can help identify spatial patterns with implications for e.g., the prediction of hydrological processes in ungauged locations. Previous studies identified spatial patterns of catchments with similar driving mechanisms of the rainfall-runoff process at the event scale by grouping catchments according to e.g., event runoff coefficients, event timescales (ratio of total event runoff to peak event runoff), or correlation coefficients between runoff characteristics and hydro-meteorological variables. However, the variability of rainfall-runoff events in a catchment over time has rarely been taken into account in previous catchment classification schemes. Here, we applied an event-based two-stage clustering approach to 378 essentially snow-free catchments with diverse physical and climate attributes in the contiguous United States. First, we clustered runoff events based on selected event runoff characteristics in each catchment into three clusters containing different event runoff shapes of short, medium, and long event timescales (catchment scale). Then, we clustered the catchments based solely on their hydro-meteorological event conditions corresponding to the three event runoff clusters and evaluated the identified catchment groups in terms of their physical and climate attributes (continental scale). As a result, we derived five groups comprising catchments with similar hydro-meteorological event conditions for the three event runoff shapes, revealing a distinct spatial pattern: In catchments dominated by a humid climate with low rainfall seasonality (number of catchments n=126), mean event rainfall intensities were primarily decisive for the clustering of event runoff into different shapes. In catchments of similar climate but larger forest cover, both mean event rainfall intensities and total event rainfall sums influenced the event clustering (n=116). In very humid regions (aridity index < 0.5) showing high rainfall seasonality (n=28), the total event rainfall sum was the only factor determining the event timescale. Furthermore, in arid lowland catchments with high rainfall seasonality (n=57), the event timescale increased with increasing antecedent soil moisture, similar to the group of arid catchments with comparably lower rainfall seasonality (n=51). Thus, we assume that in arid catchments, during dry conditions, rainfall quickly became runoff via e.g., overland flow, while during wetter conditions, slower catchment flow paths were activated (subsurface flow), leading to larger event timescales. Conversely, in humid catchments, soil water storage varied less, so rainfall characteristics themselves primarily determined the shape of runoff events. Our study demonstrates that using an event-based clustering approach results in meaningful spatial catchment groups, complementing catchment classification schemes based on long-term hydrological signatures. By focusing on the temporal variability of event runoff shapes within a catchment, regions comprising catchments with similar hydro-meteorological event conditions decisive for this variability can be identified.
How to cite: Hövel, A., Stumpp, C., Woods, R., Zheng, Y., and Stockinger, M.: Identification of catchments with similar hydro-meteorological conditions during rainfall-runoff events: An event-based clustering approach for 378 catchments in the contiguous United States, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-10711, https://doi.org/10.5194/egusphere-egu25-10711, 2025.
Streamflow droughts originate from the interplay of dominant atmospheric processes (i.e., deficits in rain or snow) and compounding conditions at the land surface, specifically evapotranspiration (E) and sub-surface storage. Furthermore, these events can last from a few weeks to multiple years. The dominant drivers of streamflow droughts depend on event type and catchment characteristics. Yet, how dominant and compounding drivers of streamflow droughts vary from monthly to multi-year events and across the landscape remains unclear. To address this knowledge gap, we use a large sample of near-natural catchments in Central Europe. For this case study, we quantify the relative contribution of (i) deficits in rain and snow and (ii) anomalies in E and S to streamflow droughts of varying duration (from monthly to multi-year). To do so, we blend data from ground-based observations, reanalysis datasets, and hydrological modelling covering the last four decades. We show that the contribution of different hydro-climatological processes to streamflow droughts varies with drought duration, onset time, and catchment characteristics. By providing a synthesis of the hydro-climatological drivers of streamflow droughts for a variety of time scales and catchment types, this study can assist regional drought monitoring and management.
How to cite: Bruno, G. and Brunner, M. I.: Dominant and compounding drivers of monthly-to-multi-year streamflow droughts in Central Europe, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-16305, https://doi.org/10.5194/egusphere-egu25-16305, 2025.
Quantification of long-term partitioning of precipitation into evaporation and runoff is a fundamental pursuit in catchment hydrology. The Budyko framework provides a theoretical basis for this and estimates the evaporative fraction based on the aridity index via the Budyko curve. However, deviations from the global-average Budyko curve suggest additional controls on precipitation partitioning beyond the aridity index. We hypothesized that root zone storage capacity (Sr), defined as maximum subsurface water accessible to vegetation roots, is a key driver of these deviations. The relationship between Sr and precipitation partitioning in the Budyko space was investigated globally across >5000 catchments. Sr was calculated using the memory method based on streamflow observations and water balance. The omega parameter (ω) from Fu’s equation was used to determine the catchment’s position in the Budyko space, reflecting precipitation partitioning. Results revealed a strong positive correlation (Spearman’s ρ=0.68) between Sr and ω globally indicating Sr as a dominant driver of precipitation partitioning. Further analysis based on Köppen-Geiger climatic zone classification revealed variations in the Sr-ω relationship, with the strongest correlations observed in cold (ρ=0.87) and Mediterranean (ρ=0.83) climates, followed by temperate (ρ=0.76), tropical (ρ=0.64) and arid climates (ρ=0.61). These findings indicate that the influence of Sr on precipitation partitioning varies across different climatic regions, with a particularly strong impact observed in cold and Mediterranean climates. This study extends prior theoretical and regional insights to a global scale confirming Sr as a governing factor in modulating catchment precipitation partitioning in the Budyko space. Vice-versa, the water and energy limits of the Budyko space provide a theoretical basis for the upper limits in Sr found in nature. The findings emphasize the variability of the Sr-ω relationship across climatic zones and underscore the importance of incorporating Sr dynamics in global water resources assessments.
How to cite: Ibrahim, M., van Oorschot, F., Coenders-Gerrits, M., van der Ent, R., and Hrachowitz, M.: Root zone storage as a key driver of catchment precipitation partitioning in the Budyko framework, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6722, https://doi.org/10.5194/egusphere-egu25-6722, 2025.
The assumption of a closed water balance in catchments is central to many hydrological applications. However, emerging evidence suggests that catchments can lose water to surrounding areas, extending their influence beyond topographic boundaries. Understanding whether a topographic catchment acts as a groundwater importer or exporter—and how these roles evolve temporally—is crucial for effective water resource management. Previous studies using the water balance approach often relied on long-term averages and single catchments, limiting insight into temporal dynamics. This study examines inter-catchment groundwater flow (IGF) using precipitation, evapotranspiration, and discharge data from 685 gauging stations across the contiguous United States for the period 1981–2020. By employing a moving window averaging approach, we quantified IGF variability across 330 subcatchments formed by neighboring gauging stations. To enhance robustness, three independent evapotranspiration datasets and three window intervals (10, 15, and 20 years) were examined. IGF was derived as a residual term in the water balance equation, where positive IGF indicates a losing catchment and vice versa. We found that more than 60% of the subcatchments exhibit clearly increasing water losses over the study period. The median value of the annual increase of IGF has been quantified up to 0.36% of annual discharge. In this contribution, we will further explore drivers of the observed spatial variability in IGF using five time-invariant catchment descriptors (e.g., topographic characteristics, distance to the coast) and three time-variant descriptors (e.g., land use changes, precipitation seasonality) to understand possible controls of groundwater interactions. Comparison against the Budyko framework and case studies informed from the existing literature will support the reliability of our findings. This study offers new insights into the dynamic behavior of IGF and its drivers in the gauged subcatchments, advancing the understanding of groundwater interactions and informing sustainable water management practices.
How to cite: Lei, C., Tarasova, L., Basso, S., J. Cohen, M., Musolff, A., and Schmidt, C.: Spatial variability and temporal changes in inter-catchment groundwater flow across the contiguous United States, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6535, https://doi.org/10.5194/egusphere-egu25-6535, 2025.
To calculate annual values for the water balance, the water year (or hydrological year) is usually used instead of the calendar year. This is done to avoid that precipitation from one year influences runoff in the following year. In snow-dominated catchments in the northern hemisphere, for example, such a carryover would occur regularly if the calendar year were used to aggregate hydrological data. To ensure that the snow melts in the year in which it fell, calculations are usually based on hydrological years that start in early fall (e.g., October 1 or November 1). In other climates, a different start of the water year is used, e.g., to ensure that it does not start in the middle of the monsoon season. Worldwide, there is a wide variation in the definition of the start date of the water year.
As the water year is used for many hydrological analyses, all annual statistics are potentially influenced by the chosen start date. The water balance for a particular year depends on how the 12-month periods is defined. Similarly, the definition of the 12-month periods is also important when calculating statistics such as annual peak flows, as it depends on whether large peaks in, for example, April and October are assigned to the same year (i.e., only one of them is considered) or to two years.
In this study, we use a modeling approach to numerically evaluate the definition of a water year and discuss how the water year would be best defined for different hydroclimatic regions. For this purpose, the runoff and storage was simulated for over 600 catchments in the USA with the HBV model. We analyzed the time series of the simulated snow, soil and groundwater storage and defined the ideal starting point of the water year from a water balance perspective as the date for which the interannual variation in total storage is the smallest. The results show that the optimal definition of the starting point of the water year varies considerably from region to region.
How to cite: Vis, M., Clerc-Schwarzenbach, F., Staudinger, M., van Meerveld, I., and Seibert, J.: What would be the best date to start the water year?, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11109, https://doi.org/10.5194/egusphere-egu25-11109, 2025.
Large-sample datasets of catchments offer opportunities to explore the hydroclimatic and physiographic controls on hydrological processes across various spatial and temporal contexts. This study leverages a global dataset of over 4,000 catchments to investigate how annual precipitation, potential evapotranspiration, and seasonal synchronicity influence streamflow dynamics. Seasonal synchronicity, reflecting the temporal alignment of precipitation and evapotranspiration, is identified as a key factor in shaping hydrological responses and improving our understanding of inter-annual variability.
We use a hybrid-modelling framework where a dense neural network, trained on catchment descriptors, is employed to parameterize a simple annual hydrological model. The hydrological model is characterized by three easily interpretable coefficients, each representing the sensitivity of annual streamflow to precipitation, evapotranspiration, and their synchronicity. By systematically evaluating regionalization across spatial, temporal, and spatiotemporal contexts, we demonstrate the potential for transferring insights and functional understanding from data-rich to data-scarce catchments.
This work contributes to advancing hydrological synthesis by linking catchment descriptors with dominant hydrological controls and exploring the representativeness of global catchment datasets. Our findings underline the importance of harmonized large-sample datasets and systematic workflows for uncovering annual hydrological processes and enabling robust predictions in ungauged basins.
How to cite: Degenne, A., Bourgin, F., Andréassian, V., and Perrin, C.: Annual Streamflow Modelling Using Large-Sample Datasets: Insights from Hybrid Models and Seasonal Synchronicity, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6732, https://doi.org/10.5194/egusphere-egu25-6732, 2025.
A large transient multi-scenario and multi-model ensemble of future streamflows and groundwater projections in France developed in a national project named Explore2 was recently published (Sauquet et al., 2024). The main objective of the Explore2 dataset is to provide a rich and spatially consistent information for the future evolution of hydrological resources in France using a large ensemble of EURO-CORDEX regional climate projections (Coppola et al., 2021) and a large variety of hydrological models.
The aim of the present study is to use a classification of river flow regimes on the hydrological projections from the Explore2 dataset to assess how hydrological processes will change in response to climate change at the catchment scale.
A simple, but well-adapted classification based on a hierarchical cluster analysis is adopted here. The classification is based on the twelve monthly Pardé coefficients derived from 611 time series of near natural observed streamflow, leading to seven characteristic river flow regimes in France over the period 1976-2005.
The Pardé coefficients were computed on 30-year periods for four time slices, namely the baseline (1976-2005), near future (2020-2049), mid-century (2041-2070), and end of the century (2070-2099) periods for each hydrological projection and for 2500 simulation points located across France. A representative regime is assigned to each simulation point corresponding to the most frequent regime identified among all hydrological projections.
River flow regime derived from the historical runs is used to assess the performance of the hydrological models at each gauged basins. The shifts in river flow regimes (between the future and baseline periods) reflect and summarize the evolution in rainfall-runoff processes due to climate change.
Overall, the predominantly rain-fed hydrological regimes will change for more contrasted regime during the 21st century. The basins with transition regimes (combining snow and rain contributions) will likely shift towards pluvial regimes. Basins at higher altitudes will keep their nival character but will have less contrasted regimes, with potentially less severe low flow in winter, and a decrease in summer flow for rivers influenced by glaciers.
References:
Coppola et al.: Assessment of the European Climate Projections as Simulated by the Large EURO-CORDEX Regional and Global Climate Model Ensemble, J. Geophys. Res.: Atmos., 126, e2019JD032356. https://doi.org/10.1029/2019JD032356, 2021.
Sauquet et al.: A large transient multi-scenario multi-model ensemble of future streamflows and groundwater projections in France, ESSD, submitted.
Strohmenger et al.: On the visual detection of non-natural records in streamflow time series: challenges and impacts, Hydrol. Earth Syst. Sci., 27, 3375–3391, https://doi.org/10.5194/hess-27-3375-2023, 2023.
How to cite: Sauquet, E., Strohmenger, L., and Thirel, G.: Impact of climate change on streamflow in France: Towards new river flow regimes?, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15892, https://doi.org/10.5194/egusphere-egu25-15892, 2025.
Europe hosts a diverse range of seasonal flow regimes, but this diversity remains only partially characterized. We use directional statistics to analyze streamflow seasonality across Europe, examining the timing and the temporal concentration of seasonal flow regimes. Geographical differences in seasonality are primarily shaped by climate, but regionally distinct imprints of landscape conditions (e.g. geology) appear. In many catchments, river flow seasonality has changed in recent decades. Seasonality has dampened in most nival flow regimes and increased in most pluvial flow regimes. Shifts in timing tend to be more variable within geographically and physiographically similar regions but can be linked to shifts in climate seasonality.
How to cite: Berghuijs, W., Carugati, S., and Hale, K.: Shifting seasonality of flow regimes across Europe, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-1991, https://doi.org/10.5194/egusphere-egu25-1991, 2025.
The study of the occurrence of abrupt hydrological extremes changes, and in particular the shift from extended dry periods to flood events is of particular interest for the stakeholders dealing with the operational management of environmental hazards. In fact, when floods occur during or soon after drought periods, early warning systems can become inaccurate and water management practices may fall short in reducing the risk of both extremes (e.g. Fowler et al., 2020; Ward et al., 2020).
In this study, we take advantage of the hydrometeorological time-series from more CAMELS-type datasets around the globe to propose and test indicators to characterise the drought-to-flood transition, understood as the basin's propensity to generate flood volumes during or immediately after droughts.
In a first phase, high flow events are identified using a fixed streamflow threshold, while drought events are characterized by means of the Standardised Precipitation Evapotranspiration Index (SPEI). Then, a set of signatures based on the co-occurrence and severity of compound drought and high flow events are calculated for all the study catchments, and their spatial pattern across the different areas of the globe is analysed. In particular, such indicators consider both the magnitude and the seasonality of high flow volumes occurring during or after drought periods.
The proposed metrics could support the hydrologic community in understanding the dynamics guiding compound events across the continents. Moreover, they could be useful in climate change impact studies to assess the evolution of combined drought and flood periods with important implications for practical risk management.
References
Fowler, K., Knoben, W., Peel, M., Peterson, T., Ryu, D., Saft, M., Seo, K., & Western, A. (2020). Many Commonly Used Rainfall-Runoff Models Lack Long, Slow Dynamics: Implications for Runoff Projections. Water Resources Research, 56(5), e2019WR025286. https://doi.org/10.1029/2019WR025286
Ward, P. J., De Ruiter, M. C., Mård, J., Schröter, K., Van Loon, A., Veldkamp, T., Von Uexkull, N., Wanders, N., AghaKouchak, A., Arnbjerg-Nielsen, K., Capewell, L., Llasat, M. C., Day, R., Dewals, B., Di Baldassarre, G., Huning, L. S., Kreibich, H., Mazzoleni, M., Savelli, E., … Wens, M. (2020). The need to integrate flood and drought disaster risk reduction strategies. Water Security, 11, 100070. https://doi.org/10.1016/j.wasec.2020.100070
How to cite: Neri, M., Selleri, G., and Toth, E.: Drought-to-flood transitions: exploring new indicators using large-sample datasets, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7023, https://doi.org/10.5194/egusphere-egu25-7023, 2025.
A key aim of large sample hydrology is to gain generalizable insights by comparing the functioning of hydrological systems across many locations and, by doing so, across many spatial gradients. To be informative, large sample datasets must therefore contain locations that cover the gradients of interest (e.g. in climate, topography, geology) and they must quantify these gradients in a meaningful way (e.g. as catchment attributes). Despite the increasing availability of open datasets with growing coverage and diversity, recent research has highlighted several limitations of the datasets currently used, which may compromise the insights gained in large sample studies. Here we will discuss three problems associated with large-sample hydrological data, as well as some possible consequences and solutions.
(1) Data uncertainty, which arises because we cannot exactly measure the variables of interest at the relevant scales.
(2) Data representativeness, i.e., the issue that our data may not represent the actual variables of interest because they are either measured indirectly, must be processed in some way, or contain subjective choices.
(3) Data imbalance, i.e., the issue that certain regions are omitted or disproportionately represented in our datasets, which may lead to biased results.
While identifying and addressing these issues is challenging, it will not only increase the value of large sample datasets, but ultimately also improve our understanding of hydrological processes and our predictive modeling capabilities.
How to cite: Gnann, S. and Wagener, T.: Large sample hydrology and the value of data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-6450, https://doi.org/10.5194/egusphere-egu25-6450, 2025.
Large-sample hydrology (LSH) datasets have advanced hydrological research by enabling studies across a wide range of catchments. Yet, the impact of landscape attributes included in such datasets on their ability to inform perceptual understanding of catchment behaviour remains underexplored. Here we investigate how the level of detail in maps used to derive catchment-scale geological attributes influences their correlation with streamflow signatures. For this, we used a set of streamflow signatures, and climate and landscape attributes available from the recently released EStreams dataset, alongside geological attributes derived from three geology maps of varying levels of detail: global, continental, and regional. These maps are perceived to have increasing levels of accuracy and were reclassified into four permeability classes. In order to explore scale-dependent effects, we moved from breadth to depth, that is, from a broad continental scale with less detailed analyses to a finer sub-catchment setting with more detailed investigations. We found that the correlation between streamflow signatures and geology attributes generally increased when using more detailed geological maps, drastically changing the perception of the importance of geology in influencing catchment behaviour relative to other landscape properties. In the Moselle catchment, a global geology map with other catchment attributes (e.g., climate and soils) failed to capture regional variations in many streamflow signatures. Moving to the sub-catchment level, we observed that smaller, nested sub-catchments exhibited unique correlation patterns, particularly for the baseflow index, emphasizing the nuanced controls at finer scales. Overall, regional and continental maps generally captured geological details better than global maps. This was particularly evident in areas with heterogeneous rock types, where global maps often oversimplified rock classifications. These findings underscore the importance of region-specific characteristics, which become even more pronounced at local scales, and play a crucial role in detecting meaningful correlations. This has implications for hydrological regionalization and predictions in ungauged catchments, suggesting that integrating high-quality, region-specific geological data into LSH studies is essential for accurate predictions and deeper insights into dominant streamflow generation processes.
How to cite: Nascimento, T., Rudlang, J., Gnann, S., Seibert, J., Hrachowitz, M., and Fenicia, F.: Geological map level of detail and its impact on our perception of dominant streamflow processes in large-sample hydrological studies, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-527, https://doi.org/10.5194/egusphere-egu25-527, 2025.
Snowmelt-driven streamflow provides water for ecosystems and one billion people. The total and temporal variability of available streamflow depends on how water fluxes are influenced by individual catchments and their climatic setting. Climate factors typically cause different seasonal and annual water balance between large regions but influences on local variations remain less understood. Here we show how both climate and landscape (expressed as soil drainage nonlinearity) control seasonal and annual water balances of 219 snowy catchments across the contiguous United States. These highly diverse catchments are first classified into three groups that are largely climatologically homogenous. This grouping indicates that climate (aridity and climate seasonality) causes distinct hydrological differences between regions. We apply a common framework to these separate group that indicates that climate also shapes what factors further drive within-region differences. Specifically, in humid catchments with winter-dominated precipitation (located in the Pacific Northwest) streamflow seasonality and annual water balances are insensitive to differences in the fraction of precipitation falling as snow (snow fraction). In relatively arid catchments with winter-dominated precipitation (located in the Rocky Mountains) larger snow fractions lead to more annual streamflow with stronger streamflow seasonality and their higher soil drainage nonlinearity enhances these effects. However, in the Northeast and the Great Lakes (where precipitation is less seasonal or summer dominated) higher soil drainage nonlinearity leads to less streamflow. We explain these paradoxical sensitivities by showing how the effect of soil drainage nonlinearity and snow fractions vary regionally depending on the prevailing water and energy balance regimes.
How to cite: Wang, Z., Berghuijs, W. R., Howden, N. J. K., and Woods, R.: Climate and landscape control on seasonal and annual water balance in snow-influenced catchments, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-3523, https://doi.org/10.5194/egusphere-egu25-3523, 2025.
The impact of climate on seasonal streamflow regimes is generally well understood in a qualitative sense, and forms the basis for classifications of streamflow regimes (e.g. pluvial, nival, glacial etc). However, these classifications are mainly descriptive, and have limited resolution since the number of classes is small. Hydrology would be further advanced if (i) we could make a quantitative synthesis using hydrological process knowledge, and (ii) we could resolve in more detail the differences between catchments which belong to the same class.
In this presentation we focus on the seasonal flow regimes of snow-influenced catchments, making use of large sample data sets such as the EStreams data as well as CAMELS data (USA, Chile). There is a substantial body of research on how climate affects the streamflow regimes of snow-affected catchments of the (western) USA, but only limited synthesis beyond this region. Here we will start to examine the extent to which lessons learned from US seasonal snow hydrology are transferable elsewhere, and explore the reasons for differences.
How to cite: Woods, R., Wang, Z., Howden, N., and Berghuijs, W.: On the Global Consistency of Climate Impacts on Seasonal River Flow Regimes in Snow-Influenced Catchments, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-19161, https://doi.org/10.5194/egusphere-egu25-19161, 2025.
Climate change is expected to modify hydrological processes, but its impacts on streamflow remain poorly understood across large geographic regions. Using streamflow and climate data from ~6,000 European catchments, we quantify the responses of streamflow means and extremes to changes in precipitation and potential evapotranspiration at annual and seasonal timescales. We find that annual mean, minimum, and maximum flows generally positively scale with mean annual precipitation, but with overall different responses. For most catchments, changes in streamflow are percentage-wise larger than those of precipitation, indicating an amplification of climate impacts on hydrology. The sensitivities of annual minimum flows are generally dampened compared to precipitation changes. We also discuss how streamflow responds to potential evapotranspiration and explore the role of catchment characteristics on these climate sensitivities of streamflow. This research helps to understand hydrological responses to climate change, which can improve water management and flood-risk mitigation across Europe.
How to cite: Hemshorn de Sánchez, A. L., Berghuijs, W., Van Loon, A., Hendriks, D., and van der Velde, Y.: Climate sensitivities of mean and extreme flows across Europe, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17310, https://doi.org/10.5194/egusphere-egu25-17310, 2025.
Previous studies show high uncertainty of evapotranspiration (ET) estimates for Norway, ranging from about 200 to more than 500 mm/year. It is partly due to sparce ET measurements in Norway to constrain the ET process in the hydrological modelling, and partly due to the uncertainty of precipitation (P) observations, especially in high mountainous regions with complex topography. In this study, we aim to quantify the uncertainty of ET estimates for 66 Norwegian catchments based on three hydrological models and four global remote sensing products. All the selected catchments are small (< 1000 km2), non-regulated and non-glacierized, with long-term discharge (Q) observations and runoff coefficient less than one (i.e. P>Q). These catchments differ substantially in climatic and hydrological characteristics and span the five hydrological regimes commonly used in Norway (Atlantic, Mountain, Inland, Baltic, and Transition). The three hydrological models (HBV, LISFLOOD and Shyft) are calibrated against daily discharge between 1981 and 2000 using three objective functions (KGE, KGE+LKGE, and KGE+BoxcoxKGE) and validated between 2001 and 2020 using six criteria (KGE, LKGE, BoxcoxKGE, percent Bias (PBias), and KGE and PBias in the snow free period). The latter two criteria are specifically selected to evaluate the model performance in simulating ET given the absence of direct ET measurements. The snow free period is identified for each catchment and year based on a daily fractional snow-covered area data at 500 m spatial resolution produced from a combined MODIS dataset (MOD10A1 and MYD10A1). The four global remote sensing estimates of ET are obtained from BESSV2, ETMonitor, PML_V2 and SSEBop_V2 datasets and they are available at fine spatial resolution (≤0.05°) and monthly time steps. The calibration and validation results show that the three hydrological models perform best using different objective functions for each hydrological regime and target variable (discharge and ET). The uncertainty ranges of annual mean ET are up to 177 mm based on all hydrological model results and remote sensing products. Using the best performing hydrological model results as the benchmark, all remote sensing products overestimate ET for the mountain regime and PML_V2 gives the best estimate for the other four hydrological regimes.
How to cite: Huang, S., Silantyeva, O., Gelati, E., Yilmaz, Y. A., Engeland, K., and Tallaksen, L. M.: Quantifying the uncertainty of evapotranspiration estimates for unregulated Norwegian catchments using multiple hydrological models and remote sensing products, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-14913, https://doi.org/10.5194/egusphere-egu25-14913, 2025.
Large-sample hydrological datasets and increasing computational power allow us to conduct modelling studies that were previously impossible. For example, it is now possible to test how different model structures affect the simulated streamflow dynamics and model performance for a variety of catchments instead of only a handful (but well-known) catchments. While this is excellent progress, the modeller generally does not understand the physical processes nor the reliability of the data for the hundreds of catchments as well as for the limited number of catchments.
The two Brazilian large-sample hydrology datasets CAMELS-BR and CABra were created for the same purpose but differ in content as they are based on different types of meteorological data. CAMELS-BR is mainly based on large-scale data from satellite, reanalysis, and gauge data, while CABra is based on the interpolation of station data. Especially the potential evapotranspiration values differ strongly for the two datasets, with the annual sums in CAMELS-BR being only 50 to 70 % of those in CABra. This situation enables us to test if different model structures enhance process representation or mainly compensate for flawed input data.
We tested the two datasets with three versions of the HBV model. Aside from the standard version simulating soil moisture and evapotranspiration, percolation, and streamflow from groundwater, we used a model version that can accommodate inter-catchment groundwater flow (i.e., inflow or outflow of groundwater), as well as a simpler version of the model in which a constant part of the precipitation is assumed to become groundwater and the remaining part is not explicitly included (i.e., soil moisture and evapotranspiration are not explicitly simulated). Although evapotranspiration is represented in a very simplified manner, this has the advantage that no (uncertain) potential evapotranspiration data are required. Interception losses can be represented better as they are not part of the evapotranspiration from the soil routine.
Our results show that large-sample datasets are very useful for testing different model structures and thus representation of hydrological processes. Regardless of the dataset used, the model version has a large effect on the streamflow simulations. The best results were usually achieved with the simplified soil routine, even though it has fewer parameters that need to be calibrated. Allowing for intercatchment groundwater flow improved the performance compared to the standard version of the model in many cases as well.
How to cite: Clerc-Schwarzenbach, F., Meyer Oliveira, A., Vis, M., Seibert, J., and van Meerveld, I.: Modelling Brazilian hydrology using various input datasets and model structures, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-2765, https://doi.org/10.5194/egusphere-egu25-2765, 2025.
Long Short-Term Memory (LSTM) networks have recently emerged as powerful data-driven approaches for rainfall-runoff modeling, often outperforming traditional hydrological models. However, their application has been predominantly tested on daily time steps and larger catchments (>250 km²). In this study, we push these boundaries by investigating the potential of LSTMs for flash flood prediction in smaller, fast-responding catchments. We leverage a refined version of the CAMELS-DE dataset, processed at hourly resolution, to capture the rapid hydrological dynamics that typify flash flood events. Hourly discharge and water level observations from federal agencies in Germany are combined with meteorological inputs from the German Weather Service (DWD), enabling a detailed assessment of the benefits of refined temporal resolution for LSTM-based modeling.
Our findings reveal that while LSTMs demonstrate reasonable skill in predicting peak discharges and event timing, performance degrades significantly during summer convective storms, characterized by localized and intense rainfall. We investigate whether this drop in performance is related to limitations in the LSTM architecture and training strategy or is due to increasing uncertainties in the meteorological boundary conditions. We further investigate when, where and how the use of hourly resolution data affects model performance. The study provides critical insights into the challenges and opportunities of using data-driven approaches for flash flood forecasting in small, fast-responding catchments, contributing to the development of more robust hydrological prediction systems. In addition, we present a preliminary version of CAMELS-DE in hourly resolution, opening new possibilities for research in the field of large sample hydrology.
How to cite: Dolich, A., Acuña Espinoza, E., and Loritz, R.: Towards Accurate Flood Predictions in Small, Fast-Responding Catchments Using Hourly CAMELS-DE Data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-11922, https://doi.org/10.5194/egusphere-egu25-11922, 2025.
The increasing variability of precipitation and temperature extremes under climate change requires advanced methodologies to better understand and predict watershed responses. Watersheds with similar features are expected to exhibit comparable hydrological responses to meteorological events, and by clustering them, we aim to improve knowledge transfer from data-rich to data-scarce regions and enhance hydrological process analysis and prediction.
Our approach leverages machine learning to cluster watersheds based on shared characteristics, such as topography, land cover, soil properties and geology, proposing an expanded perspective on watershed similarities and their implications on the understanding of hydrological phenomena.
We utilize 62 lumped watershed descriptors provided by the LamaH-CE large-sample hydrology dataset (https://doi.org/10.5194/essd-13-4529-2021) including key attributes for each catchment, such as area, mean elevation, slope, land use, NDVI, soil porosity, and rock permeability. A Principal Component Analysis (PCA) was first applied to reduce dimensionality and identify the most significant watershed descriptors. Next, four unsupervised learning models - K-means, Gaussian Mixture Models (GMM), Hierarchical Clustering, and DB Scan - were implemented for clustering the watersheds using the selected descriptors. The models’ performances were systematically evaluated and compared regarding shape factors and cluster interpretation across different watershed categories. Advanced dimensionality reduction techniques and arbitrary descriptor selection were tested to ensure robustness of the procedures. Stability testing and hyperparameter optimization further confirmed the clustering models. The resulting clusters were explored through detailed maps and 2D and 3D plots, revealing patterns of similarity across diverse geographic and hydrological regions in the LamaH-CE domain. For instance, watersheds that are characterized by large areas and modest elevations ranges are in the same cluster, even if they are not hydrologically connected or close to each other. Especially when working at large spatial scales, where basins with different response types are analysed together, watershed clustering allows to tailor specific modelling and analysis techniques for different watershed clusters, providing additional and more precise knowledge on watershed behaviour.
Future research steps will focus on testing this methodology as a basis for transferring knowledge from gauged to ungauged basins within the same cluster, enhancing predictive capabilities in data-scarce regions.
Beyond hydrological predictions, the clusters of watershed characteristics can also find applications in water resources planning and management in low-data regions supporting more informed decision-making.
How to cite: Bertoli, G., Schröter, K., Arcucci, R., and Caporali, E.: Exploring watershed similarities through machine learning and watershed descriptors: enhancing hydrological predictions, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-17403, https://doi.org/10.5194/egusphere-egu25-17403, 2025.
One of the current goals of large-scale hydrological modelling within the FOCCUS project (https://foccus-project.eu) is improved estimation of water and matter runoff to the coastal regions of Europe. The main challenges and decisions that are typically made for physically based modelling are as follows: the complexity of the model, the level of process representation in the model, the number of models involved in the modelling chain, and a level of regionalization of model parameters. Creating a parsimonious model and using enough data to achieve the desired results are also challenging. In order to account for geographical variability of the model parameters, calibration at a large number of subbasins is necessary when implementing increasingly complex hydrological models at large scale. Our research focuses on the level of regionalization of parameters and assessment of model performance for coastal vs full domain calibration.
The Hydrological Predictions for Environment (HYPE) model (1), used here for large-scale hydrological analyses, is a dynamic process-based rainfall-runoff and water quality model. Regional precipitation corrections tested in the previous European-domain HYPE (E-HYPE) version improved the model performance by helping to compensate for systematic biases in the water balance (2). However, the addition of these precipitation corrections also introduces uncertainty to the model calibration, as the corrections could just compensate for model biases instead of addressing underlying problems in the model structure or process representation. The calibration strategy for the current E-HYPE version 4 has one “global” parameter set that is used for the European domain to limit the amount of parameter regionalization. E-HYPE calibration was further improved using performance filters for parameter sets and assessment of model behaviour including snow water equivalent, reservoir sedimentation, and stream resuspension. We explore the effects of regional calibration against the coastal gages on the overall model performance and process description as well as on the freshwater inflows to coastal areas. We discuss the trade-offs between regionalised coastal-oriented calibration and full-domain model tuning, point out the main factors limiting model performance, and investigate the effect of diversified calibration workflow based on soil/landuse dependent parameters.
References:
- SMHI, 2023. HYPE Model Documentation. 〈http://www.smhi.net/hype/wiki/〉
- Brendel, C., Capell, R., & Bartosova, A. (2023). To tame a land: Limiting factors in model performance for the multi-objective calibration of a pan-European, semi-distributed hydrological model for discharge and sediments. Journal of Hydrology: Regional Studies, 50, 101544.
How to cite: Terskii, P., Bartosova, A., and Brendel, C.: Trade-offs between regionalised and non-regionalized calibration of a continental model for estimating European coastal river inflow, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-15548, https://doi.org/10.5194/egusphere-egu25-15548, 2025.
Land/hydrologic model advances have significantly enhanced the capability to simulate complex hydrological processes. However, the accuracy of these simulations is often undermined by uncertainties in model parameters, many of which are poorly constrained by observations, as well as the high computation demand of sophisticated models, which restricts their optimization. To address these challenges, we developed a machine learning (ML)-based calibration approach using large-sample emulators (LSEs) to optimize and regionalize model parameters. We have now evaluated an LSE for three models spanning a range of complexity: the conceptual HBV hydrologic model, the process-based SUMMA hydrologic model, and the Community Terrestrial Systems Model (CTSM) land model.
Our LSE approach leverages static catchment attributes and parameter values across basins to train ML emulators, which are then coupled with optimization algorithms (e.g., Genetic Algorithm) to iteratively refine parameter estimates. This iterative process enhances both the accuracy and number of parameter samples, progressively improving model performance. Results show that the LSE-based optimization achieved median modified Kling Gupta Efficiency (KGE') values of 0.65 for CTSM, 0.76 for SUMMA, and above 0.8 for HBV. Those values are competitive with or better than comparable model calibration results in past years, and outperform local calibrations based on single-site emulators (SSEs). This presentation will highlight the methodologies, key results, and challenges of implementing LSE-based calibration for hydrologic and land models at a continental scale, emphasizing the potential for regionalization and improved predictive capabilities in large-domain hydrologic modeling.
How to cite: Tang, G., Wood, A., Farahani, M., Mizukami, N., and Swenson, S.: Advancing continental-scale hydrology model calibration using large-sample emulators: from simple conceptual to complex process-based land models , EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-7142, https://doi.org/10.5194/egusphere-egu25-7142, 2025.
In the early 1900s, engineer Carl E. Grunsky developed an empirical equation relating mean annual precipitation to mean annual runoff in the San Francisco Peninsula, California, United States of America (USA). Over a century later, this method, now referred to as the R-Grunsky, was reintroduced and generalized by incorporating a rainfall-streamflow coefficient (α) derived from mean annual temperature and precipitation. Previous studies have demonstrated its effectiveness in Mediterranean and Brazilian catchments. Here, we aim to expand the applicability of the R-Grunsky method by analyzing data from 14,894 catchments from the Caravan dataset, spanning Australia, Brazil, Chile, Great Britain, Europe, and the USA. We first developed a global relationship between α and the catchments' mean temperature and precipitation, subsequently fitting a multiple linear regression model to estimate α values for streamflow prediction using a simplified equation system. The R-Grunsky approach showed suitable results in mean annual streamflow estimation, with a Kling-Gupta Efficiency (KGE) of 0.47 and R² = 0.66 considering all studied catchments. The results align with those from previous studies of the method developed to Brazilian catchments. However, we noted that arid regions, primarily in Australia, central USA, northeastern Brazil, and northern Chile, exhibited lower KGE values, indicating reduced performance compared to wetter catchments, which requires further investigation. The R-Grunsky approach offers the advantage of requiring only precipitation and mean temperature data for streamflow predictions, making it particularly valuable for ungauged catchments. Consequently, it holds significant potential for diverse water resource projects and decision-making processes related to water security and management.
How to cite: Marchezepe, B., Almagro, A., Ballarin, A., and Oliveira, P.: R-Grunsky: an empirical method for globally predicting streamflow using precipitation and temperature data, EGU General Assembly 2025, Vienna, Austria, 27 Apr–2 May 2025, EGU25-13788, https://doi.org/10.5194/egusphere-egu25-13788, 2025.
